Skip to content

At-risk students

Find At-Risk Students in Week 3, Not Week 15

By the time a student fails, everyone can see it. The information existed weeks earlier — a fortnight of missed classes, two skipped submissions, a mark that slid. It was just sitting in four systems that never spoke to each other.

Request a College Demo Higher Education overview

What this actually means

At risk is a definition you choose, not a verdict the software issues

No single signal identifies a struggling student. Attendance alone flags the commuter. Marks alone flag the student who had one bad week. Early identification is what happens when the signals are read together, against a threshold the institution sets and can change.

You set the rule. A department decides what counts: how many missed sessions, how far a mark has to fall, how many submissions can slip. The thresholds are visible and adjustable, not buried in a model nobody can inspect.

The flag is a prompt, not a label. Nothing is written on a student's record. A flag reaches an advisor as a suggestion that a conversation is due, and the advisor is the one who decides whether it is warranted.

A flag with no owner is noise. Every raised flag goes to a named person with a suggested action. An alert that lands in a shared inbox is a report, and reports do not change outcomes.

The signals

Five things that show up before a result does

Each of these is weak on its own and meaningful in combination. The platform watches them together and only raises a flag when the combination crosses the threshold the institution set.

  • Signal 01

    Attendance that changes shape

    Not the absolute figure, but the break in pattern: a student who attended everything and then stopped attending Tuesdays. The change is the signal.

  • Signal 02

    A falling trend, not a low mark

    A consistent 55 is a known quantity. A student moving from 78 to 61 to 52 across three assessments is a different situation and deserves a different response.

  • Signal 03

    Submissions quietly skipped

    Missed and late submissions usually precede a bad result by weeks, and they rarely reach anyone who could intervene until marks are collated.

  • Signal 04

    A gap that compounds

    A concept that never got fixed and is now a prerequisite for the current unit. This student is not lazy; they are building on something that was never solid.

  • Signal 05

    Disengagement from practice

    A student who stops opening material, stops attempting practice and stops asking the tutor anything has usually decided something before anyone has noticed.

A worked example

What actually happens after a flag is raised

Identification is the easy half. The reason early-warning projects fail is that nobody defined what happens next, so the list gets produced, circulated and ignored. This is the loop the platform is built around.

  1. The flag is raised

    Signals cross the department's threshold in week three. The flag lists what triggered it, so the advisor knows the reason before they open the file.

  2. A person looks at it

    The advisor or mentor confirms, dismisses or escalates. Dismissals are recorded too, which is how the thresholds get better over a couple of terms.

  3. Something specific happens

    A remedial path on the concept that broke, a mentor meeting, or a referral. The action is attached to the flag rather than remembered by whoever made it.

  4. It is measured again

    The same concept is re-assessed after the intervention. Either the student moved or they did not, and the second number is what tells you which.

Step four is the one institutions skip, and it is the one that turns an early-warning list into a practice. See the metrics these flags are built from.

Questions

What student support teams ask

Is this a prediction about whether a student will fail?

No, and we are careful about that distinction. The platform surfaces students whose observable signals have crossed a threshold the institution defined. It does not assign a probability of failure or a risk score that follows a student around. The judgement about what the signals mean stays with the advisor.

Does a student know they have been flagged?

That is the institution's policy, not a platform default. Many colleges prefer the flag to stay with the advisor and the conversation to happen without the label, and the system supports that. Whichever way a college chooses, flags are visible to the staff responsible for the student, not to their peers.

How do we avoid flagging half the cohort in week two?

By starting with conservative thresholds and tuning them. Most departments begin with a narrow definition so the first list is small enough to act on, then widen it once the loop is working. Dismissed flags are recorded, which gives you the evidence to adjust rather than guess.

What if attendance is recorded outside the platform?

It usually is. Attendance, enrolment and programme structure sync from the academic ERP, so the signals combine without anyone re-entering data. Where a college has no ERP feed, attendance can be captured on the platform directly.

Define at risk for one department and see who appears

Pick a programme, tell us what your team already counts as a warning sign, and we will show you what the flag list would look like and who it would go to.

Request a College Demo See the analytics behind it

Related